An official website of the United States government
In many real-world problems, field measurements are limited and exhibit non-stationarity across locations, making it difficult to build reliable predictive models using field data alone. To address data limitations, simulation models are often employed to provide broader spatial coverage. However, data from different sources often fail to align because they are collected at different spatial scales. We propose a framework for integrating spatial data collected at different scales through learning two interdependent spatial models. At the coarse scale, a Multi-Output Gaussian Process (MOGP) jointly models coarse-scale simulation data and aggregated field measurements as correlated outputs of a shared latent spatial process. This model enables calibration of simulation outputs using sparse field observations at coarse resolution. At the fine scale, a Deep Gaussian Process (DGP) captures non-stationary spatial patterns in the field data. To ensure consistency across spatial scales, the DGP’s fine-level predictions are constrained to match the coarse-level totals calibrated by the MOGP. In addition, the DGP iteratively improves the accuracy of aggregated field measurement. A key innovation of our approach is the iterative, joint training of the two models, i.e., the MOGP provides top-down guidance by imposing coarse-level constraints on the DGP, while the DGP offers bottom-up feedback by aggregating its fine-scale predictions and using them to refine the MOGP estimates. This mutual reinforcement improves both the accuracy and coherence of predictions across scales. We validate the proposed framework on a case study of hurricane debris estimation in Florida, demonstrating improvements in predictive performance and uncertainty quantification over existing methods.
more »
« less
Multi-Scale Graph Learning for Anti-Sparse Downscaling
Water temperature can vary substantially even across short distances within the same sub-watershed. Accurate prediction of stream water temperature at fine spatial resolutions (i.e., fine scales, ≤ 1 km) enables precise interventions to maintain water quality and protect aquatic habitats. Although spatiotemporal models have made substantial progress in spatially coarse time series modeling, challenges persist in predicting at fine spatial scales due to the lack of data at that scale. To address the problem of insufficient fine-scale data, we propose a Multi-Scale Graph Learning (MSGL) method. This method employs a multi-task learning framework where coarse-scale graph learning, bolstered by larger datasets, simultaneously enhances fine-scale graph learning. Although existing multi-scale or multi-resolution methods integrate data from different spatial scales, they often overlook the spatial correspondences across graph structures at various scales. To address this, our MSGL introduces an additional learning task, cross-scale interpolation learning, which leverages the hydrological connectedness of stream locations across coarse- and fine-scale graphs to establish cross-scale connections, thereby enhancing overall model performance. Furthermore, we have broken free from the mindset that multi-scale learning is limited to synchronous training by proposing an Asynchronous Multi-Scale Graph Learning method (ASYNC-MSGL). Extensive experiments demonstrate the state-of-the-art performance of our method for anti-sparse downscaling of daily stream temperatures in the Delaware River Basin, USA, highlighting its potential utility for water resources monitoring and management.
more »
« less
- PAR ID:
- 10584673
- Publisher / Repository:
- AAAI
- Date Published:
- Journal Name:
- Proceedings of the AAAI Conference on Artificial Intelligence
- Volume:
- 39
- Issue:
- 27
- ISSN:
- 2159-5399
- Page Range / eLocation ID:
- 27969 to 27977
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
Thermal Regimes of Beaver‐Impacted Stream Reaches in the Tualatin River Basin, Oregon, United StatesABSTRACT As beavers are increasingly used in stream restoration, interest in their potential to mitigate stream degradation has grown. However, the impacts of beaver dams on stream temperature remain inconsistent across studies, often due to differences in spatial scale and methodology. We investigated summer thermal regimes at multiple spatial scales in three beaver‐impacted streams within the Tualatin River Basin, Oregon, using fiber‐optic distributed temperature sensing. Substantial variability in thermal regimes was observed among the three stream reaches and around individual beaver dams within each reach. This variability was primarily influenced by geomorphic changes in stream depth associated with beaver dams and restoration history. The thermal sensitivity of stream temperatures to air temperatures significantly increased across the three sites, following a gradient of increasing time since restoration, beaver activity, and stream depth. Across all sites, stream depth was more strongly correlated with variability in daily maximum temperatures than canopy cover. This was most evident at Fanno Creek, where stream depth explained ~41% of the variability in daily maximum temperatures on August 16, 2023. Reach‐scale measurements often masked fine‐scale thermal heterogeneity with ecological implications. At the beaver dam scale, significant differences in daily maximum temperature and daily temperature range were observed among upstream, pond, and downstream locations, emphasizing the thermal variability induced by beaver dams over short distances. Our findings demonstrate that thermal regimes in beaver‐impacted reaches are shaped by local geomorphic factors and vary with spatial scale, underscoring the importance of multi‐scale approaches to assess the thermal effects of beaver activity.more » « less
-
Accurate prediction of water quality and quantity is crucial for sustainable development and human well-being. However, existing data-driven methods often suffer from spatial biases in model performance due to heterogeneous data, limited observations, and noisy sensor data. To overcome these challenges, we propose Fair-Graph, a novel graph-based recurrent neural network that leverages interrelated knowledge from multiple rivers to predict water flow and temperature within large-scale stream networks. Additionally, we introduce node-specific graph masks for information aggregation and adaptation to enhance prediction over heterogeneous river segments. To reduce performance disparities across river segments, we introduce a centralized coordination strategy that adjusts training priorities for segments. We evaluate the prediction of water temperature within the Delaware River Basin, and the prediction of streamflow using simulated data from U.S. National Water Model in the Houston River network. The results showcase improvements in predictive performance and highlight the proposed model's ability to maintain spatial fairness over different river segments.more » « less
-
This paper introduces a novel approach to fine-tuning transformer-based models for various spatial-temporal downstream tasks. While fine-tuning approaches have shown remarkable success in fields like natural language processing, their efficacy in human-generated spatial-temporal data is often hindered by the complex spatial-temporal correlation and extensive reliance on labeled data. We introduce Knowledge Graph-Guided Spatial-Temporal Cross-task Fine-Tuning method, i.e., KG-STFT, a cross-task fine-tuning approach designed for adapting to data-scarce spatial-temporal tasks by leveraging rich knowledge embedded in similar downstream tasks and pre-trained model. Our KG-STFT framework utilizes i) a non-linear knowledge ensembler to capture and integrate knowledge embedded in different transformer blocks, and ii) constructs a task knowledge graph to "transfer" knowledge from data-rich to data-scarce tasks. Empirical experiments on real-world taxi trajectory data show that KG-STFT outperforms baselines, especially in data-scarce tasks, by leveraging task commonalities to improve fine-tuning.more » « less
-
Graph Neural Networks (GNNs) have become powerful tools for learning from graph-structured data, finding applications across diverse domains. However, as graph sizes and connectivity increase, standard GNN training methods face significant computational and memory challenges, limiting their scalability and efficiency. In this paper, we present a novel framework for efficient multiscale training of GNNs. Our approach leverages hierarchical graph representations and subgraphs, enabling the integration of information across multiple scales and resolutions. By utilizing coarser graph abstractions and subgraphs, each with fewer nodes and edges, we significantly reduce computational overhead during training. Building on this framework, we propose a suite of scalable training strategies, including coarse-to-fine learning, subgraph-to-full-graph transfer, and multiscale gradient computation. We also provide some theoretical analysis of our methods and demonstrate their effectiveness across various datasets and learning tasks. Our results show that multiscale training can substantially accelerate GNN training for large-scale problems while maintaining, or even improving, predictive performance.more » « less
- Free Publicly Accessible Full Text
-
Accepted Manuscript
- Journal Article:
- https://doi.org/10.1609/aaai.v39i27.35014
